M.S. Computer Science · University of Massachusetts Dartmouth
Full-Stack Engineer · AI/ML Systems · Autonomous Agents

• 🎓 M.S. in Computer Science at University of Massachusetts Dartmouth (May 2026)
• 🏛️ B.Tech in Computer Engineering — LJ University, Ahmedabad, India (2020–2024)
• 📍 Based in Dartmouth, MA — Open to relocation
• 🤖 Passionate about building AI-powered full-stack systems and autonomous agents
• 🔬 Research interests: LLM applications, deepfake detection, browser automation, ATS systems
• 💼 Currently seeking Software Engineer / AI Engineer internship opportunities
• 🌐 Portfolio: https://voraharsh.dev

Languages

Frameworks & Libraries

Cloud & Databases

AI / ML & Tools

CIS 600 Master's Project · UMass Dartmouth · 2025–2026

Full-stack AI platform that automates the end-to-end job application pipeline — from live job discovery to autonomous LinkedIn Easy Apply submission.

• 🔍 Live job scraping via Apify LinkedIn Scraper with BeautifulSoup fallback
• 🧠 Hybrid ATS Resume Scoring Engine — Python TF-IDF + synonym normalization + Groq Llama 3.1 feedback · 91.3% precision on 15-JD test set
• ✍️ AI Resume Optimizer — generates role-specific, ATS-tailored resumes from user's uploaded PDF using Groq Llama 3.1 (1,400-token budget)
• 🤖 LinkedIn Easy Apply Agent — Playwright-based Chromium automation with persistent session thread architecture, eliminating cross-thread browser crashes · 0% crash rate across 40 tested sessions
• 📊 Application Tracker — stage management (Applied → Interview → Offer → Rejected) with Gmail cross-referencing and CSV export

Stack: React Vite FastAPI Python Supabase Groq API Playwright Apify Vercel Render

2024 – Present

Dual-modality AI architecture for detecting AI-generated disinformation across text and visual modalities.

• Engineered T5 (semantic text analysis) + Swin Transformer (high-resolution visual analysis) cross-attention architecture
• Achieved 93.7% detection accuracy — +14.2% over baselines
• Integrated Explainable AI tools (LIME & Saliency Maps) for transparency and interpretability
• Curated and annotated a dataset of 8,500 manipulated and authentic media items
• Reduced false positives by 18.6% · boosted user verification correctness by 27.5%
• 🏅 3rd Runner-Up — Sigma Xi Scientific Research Honor Society

Stack: Python T5 Swin Transformer LIME PyTorch Hugging Face

2025

Secure, full-stack personal finance application with AI-driven transaction categorization and budget forecasting.

• MERN stack architecture with Groq AI for high-accuracy transaction categorization and predictive budget forecasting
• Real-time dynamic analytics dashboards for personal wealth and expense tracking
• Robust RESTful APIs with OAuth authentication for data privacy

Stack: MongoDB Express.js React Node.js Groq AI OAuth

2024

Scalable social media platform with an AI-powered recommendation engine for food security initiatives.

• AI recommendation engine combining collaborative and content-based filtering — increased user retention by 5%
• Redux state management across dynamic trend-analysis dashboards
• Node.js + MongoDB backend with efficient NoSQL schema design

Stack: React Redux Node.js MongoDB Express.js

Software Developer — Agile Infoways (Jan 2024 – Jun 2024)

• Architected and deployed a scalable project management web app using the MERN stack with RESTful APIs
• Engineered a Role-Based Access Control (RBAC) system with JWT authentication across dev/test/prod environments
• Optimized cloud-based pipelines and CI/CD workflows on AWS (EC2, S3) — reduced project delivery delays by 25%

Frontend Developer — Plushvie (Jan 2022 – Jan 2024)

• Built a real-time anonymous chat platform using Python, Django, and MySQL — supported 1,000+ concurrent users with zero downtime
• Implemented WebRTC and WebSocket protocols achieving a consistent 300ms response time
• Designed a real-time analytics dashboard with SCSS and Firebase — earned Best Team Lead Award

Relevant Coursework: Artificial Intelligence (A+) · Advanced Data Mining (B+) · Machine Learning · Algorithms

• 📧 Email: hvora@umassd.edu
• 💼 LinkedIn: linkedin.com/in/harshvora14
• 🌐 Portfolio: dev-harshvora.netlify.app
• 💻 GitHub: github.com/harshvora

Open to full-time opportunities, research collaborations, and internships in AI/ML and full-stack engineering.
Let's build something impactful together.

M.S. Computer Science · University of Massachusetts Dartmouth
Full-Stack Engineer · AI/ML Systems · Autonomous Agents

• 🎓 Pursuing M.S. in Computer Science at University of Massachusetts Dartmouth (May 2026)
• 🏛️ B.Tech in Computer Engineering — LJ University, Ahmedabad, India (2020–2024)
• 📍 Based in Dartmouth, MA — Open to relocation
• 🤖 Passionate about building AI-powered full-stack systems and autonomous agents
• 🔬 Research interests: LLM applications, deepfake detection, browser automation, ATS systems
• 💼 Currently seeking Software Engineer / AI Engineer internship opportunities
• 🌐 Portfolio: https://voraharsh.dev

Languages

Frameworks & Libraries

Cloud & Databases

AI / ML & Tools

CIS 600 Master's Project · UMass Dartmouth · 2025–2026

Full-stack AI platform that automates the end-to-end job application pipeline — from live job discovery to autonomous LinkedIn Easy Apply submission.

• 🔍 Live job scraping via Apify LinkedIn Scraper with BeautifulSoup fallback
• 🧠 Hybrid ATS Resume Scoring Engine — Python TF-IDF + synonym normalization + Groq Llama 3.1 feedback · 91.3% precision on 15-JD test set
• ✍️ AI Resume Optimizer — generates role-specific, ATS-tailored resumes from user's uploaded PDF using Groq Llama 3.1 (1,400-token budget)
• 🤖 LinkedIn Easy Apply Agent — Playwright-based Chromium automation with persistent session thread architecture, eliminating cross-thread browser crashes · 0% crash rate across 40 tested sessions
• 📊 Application Tracker — stage management (Applied → Interview → Offer → Rejected) with Gmail cross-referencing and CSV export

Stack: React Vite FastAPI Python Supabase Groq API Playwright Apify Vercel Render

2024 – Present

Dual-modality AI architecture for detecting AI-generated disinformation across text and visual modalities.

• Engineered T5 (semantic text analysis) + Swin Transformer (high-resolution visual analysis) cross-attention architecture
• Achieved 93.7% detection accuracy — +14.2% over baselines
• Integrated Explainable AI tools (LIME & Saliency Maps) for transparency and interpretability
• Curated and annotated a dataset of 8,500 manipulated and authentic media items
• Reduced false positives by 18.6% · boosted user verification correctness by 27.5%
• 🏅 3rd Runner-Up — Sigma Xi Scientific Research Honor Society

Stack: Python T5 Swin Transformer LIME PyTorch Hugging Face

2025

Secure, full-stack personal finance application with AI-driven transaction categorization and budget forecasting.

• MERN stack architecture with Groq AI for high-accuracy transaction categorization and predictive budget forecasting
• Real-time dynamic analytics dashboards for personal wealth and expense tracking
• Robust RESTful APIs with OAuth authentication for data privacy

Stack: MongoDB Express.js React Node.js Groq AI OAuth

2024

Scalable social media platform with an AI-powered recommendation engine for food security initiatives.

• AI recommendation engine combining collaborative and content-based filtering — increased user retention by 5%
• Redux state management across dynamic trend-analysis dashboards
• Node.js + MongoDB backend with efficient NoSQL schema design

Stack: React Redux Node.js MongoDB Express.js

Software Developer — Agile Infoways (Jan 2024 – Jun 2024)

• Architected and deployed a scalable project management web app using the MERN stack with RESTful APIs
• Engineered a Role-Based Access Control (RBAC) system with JWT authentication across dev/test/prod environments
• Optimized cloud-based pipelines and CI/CD workflows on AWS (EC2, S3) — reduced project delivery delays by 25%

Frontend Developer — Plushvie (Jan 2022 – Jan 2024)

• Built a real-time anonymous chat platform using Python, Django, and MySQL — supported 1,000+ concurrent users with zero downtime
• Implemented WebRTC and WebSocket protocols achieving a consistent 300ms response time
• Designed a real-time analytics dashboard with SCSS and Firebase — earned Best Team Lead Award

Relevant Coursework: Artificial Intelligence (A+) · Advanced Data Mining (B+) · Machine Learning · Algorithms

• 📧 Email: hvora@umassd.edu
• 💼 LinkedIn: linkedin.com/in/harshvora14
• 🌐 Portfolio: https://voraharsh.dev
• 💻 GitHub: github.com/harshvora

Open to full-time opportunities, research collaborations, and internships in AI/ML and full-stack engineering.
Let's build something impactful together.